In the winter time we refill the bird feeders regularly. Invariably I'd get bird seed all over the floor as I'd try to refill the container from a large bag of seed, so I decided to resolve the issue once and for all, and have some 3d printing fun along the way!
The solution was a funnel with a wide exit, and a shoulder that would enable it to sit securely on the top of the container being filled.
Didn't spill a single seed!
I drew the funnel in Sketchup, using a simple construction method, first drawing the cross-section of the desired shape and then using the "follow-me" tool to rotate that shape about a vertical axis. I've also been playing with various screen-recording tools, to find ways of enhancing blog presentations. As an experiment I've recorded the Sketchup drawing technique in "screencast-o-matic" and shared it below.
Finally... The birds were very happy with the refilled feeders!
It's a great illusion of impossibly stacked little cubes, and readily printable, believe it or not! The STL file and full detail is posted here on Thingiverse. Credit goes to Jonathan Wong for explaining how he constructed the printable version of this illusion and providing instruction on how you can draw your own version. Some additional reference to the famous triangle may also be found here.
The printing challenge: This is a challenging print in a number of respects. The nine "inverted pyramid shapes" have a very small area of contact with the print bed and don't actually support each other until over halfway built. You need to have laid down a good first layer, well bonded to the print bed. Even then, what may happen is the print head will clip any slightly curling edge as it moves about, and knock one piece over, destroying the entire print. However, there is one version of the STL file in the Thingiverse share that has a bracing piece between all the pyramids to give greater support on the print bed.
Above is what happened when my print head clipped off a slightly curled edge of one of the pyramids. I had to abandon this attempt.
For my second attempt I used a Z-Lift feature in Slic3r (the STL slicing package) which raises the print head by a prescribed amount as it moves between printing sections, lowering it again to resume printing. (Z axis lift is discussed towards the end of this older post here by RichRap.) This feature worked ok but as the object grew there seemed to be degradation of print quality with a lot of gapping in the print, even though I had only changed one parameter i.e. the z-lift. My observation was that this z-lift may have been allowing a little bit more time for retraction to take place and plastic would not resume extruding until a good length of print path had been travelled. While my retraction settings may not have been 100% initially, this setting seemed to exaggerate the problem.
Rather than spend more time examining g-code and tweaking settings I said I'd give Kisslicer a go at slicing this object. I know from past use that Kisslicer prints sub-sections to completion before moving on to the next sub-section on the same layer; meaning it prints perimeters then in-fills that section before moving on. In contrast, Slic3r prints perimeters then returns later to infill after completing all perimeters. With an object like this, with its nine base triangles, Slic3r does a lot of zipping around in contrast to Kisslicer. My hope was that Kisslicer's approach to printing would decrease the risk of knocking one of the pyramids off its base, and the less zipping about would result in less retraction movements, maintaining print quality.
Kisslicer actually printed the object to completion first time (above photo), but my choice of settings produced some quality issues also. Kisslicer has different Retraction terminology, with what it calls Destringing (Suck & Prime), a whole other evening of experimenting! The "Destringing" settings I had, resulted in quite a scraggy print and left me with a bit of clean-up to do, quite the opposite of the gapped print Slic3r produced with too much retraction.
Finally, here's a short video clip showing it rotating in and out of its illusion position.
Tech Notes and Conclusions:
The object was printed in ABS plastic with a layer height of .3mm and a layer width of .45mm. It was printed on a RepRap style 3d printer, with 3mm filament fed to a .5mm j-head hotend, on a heated bed. Gentle cooling by a low powered fan was used across the object once it climbed off the print bed. This is not my typical practice when printing in ABS (usually no fan), but helped strengthen the narrow walls as they grew.
This object has very little surface area in contact with the print bed, has many thin walls growing to points, and leaning at angles. At its default size it's a challenging print with a .5mm print head nozzle. It might be worth scaling up to a larger size or else using a finer nozzle if you have one.
I'd be interested to learn how others may have got on printing this object. If you have any tips or even quentions then feel free to add a comment below.
The Earbud Holder from Thingiverse caught my attention recently. There was a comment asking if it could have a key chain loop added... it sure can! It's a matter of personal preference whether you print the version with or without the keyring tab, or take the design further. It's great how designs evolve.
I fitted a small ring and short length of cord to the holder.
Notes:
I dropped the original Core .stl into Sketchup and drew on a tab with a hole in it, to which you could fit a keyring. I then took the evolving cover design from here: http://www.thingiverse.com/thing:38292 and put a notch in each cover so it would still close around the new keyring tab.
I then exported the new .stl files and ran them through Netfabb Cloud Service . It always does a good job of cleaning up the stl. They were now ready for printing... ABS, 50mm/sec, .4mm nozzle, .3mm width /.45mm height.
My first set of covers cracked as I tried to clip them on to the core. They broke along the print lines. On examination it was evident there was poor bonding between layers. I've been pushing the speed in recent times and at this point I expect the plastic just wasn't hot enough for the speed (230Deg C at 50mm/sec).
I increased the temperature to 245 Deg C. and reprinted just the two covers. Now, you know you should never change more than one parameter when troubleshooting, but I couldn't resist altering the layer height/width also, to increase the horizontal resolution and give a better fit between the parts. The covers were reprinted at .25mm height/.375mm width, 70mm/sec. 245 Dec C.
I've uploaded the modified design to Thingiverse: http://www.thingiverse.com/thing:39767
The solution was a funnel with a wide exit, and a shoulder that would enable it to sit securely on the top of the container being filled.
Didn't spill a single seed!
